Interplay between the Smc5/6 complex and the Mph1helicase in recombinational repairYu-Hung Chena,b, Koyi Choia,b, Barnabas Szakalc, Jacqueline Arenza,1, Xinyuan Duand, Hong Yed, Dana Branzeic,and Xiaolan Zhaoa,2

aMolecular Biology Program, Memorial Sloan-Kettering Cancer Center, New York, NY 10065; bPrograms in Biochemistry, Cell, and Molecular Biology, WeillGraduate School of Medical Sciences of Cornell University, New York, NY 10021; cItalian Foundation for Cancer Research, Institute of Molecular Oncology,20139 Milan, Italy; and dJames Graham Brown Cancer Center, University of Louisville, School of Medicine, Louisville, KY 40202

Edited by Thomas D. Petes, Duke University Medical Center, Durham, NC, and approved October 14, 2009 (received for review July 23, 2009)

The evolutionarily conserved Smc5/6 complex is implicated inrecombinational repair, but its function in this process has beenelusive. Here we report that the budding yeast Smc5/6 complexdirectly binds to the DNA helicase Mph1. Mph1 and its helicaseactivity define a replication-associated recombination subpath-way. We show that this pathway is toxic when the Smc5/6 complexis defective, because mph1� and its helicase mutations suppressmultiple defects in mutants of the Smc5/6 complex, including theirsensitivity to replication-blocking agents, growth defects, andinefficient chromatid separation, whereas MPH1 overexpressionexacerbates some of these defects. We further demonstrate thatMph1 and its helicase activity are largely responsible for theaccumulation of potentially deleterious recombination intermedi-ates in mutants of the Smc5/6 complex. We also present evidencethat mph1� does not alleviate sensitivity to DNA damage or theaccumulation of recombination intermediates in cells lacking Sgs1,which is thought to function together with the Smc5/6 complex.Thus, our results reveal a function of the Smc5/6 complex in theMph1-dependent recombinational subpathway that is distinctfrom Sgs1. We suggest that the Smc5/6 complex can counteract/modulate a pro-recombinogenic function of Mph1 or facilitate theresolution of recombination structures generated by Mph1.

The Smc5/6 complex is one of the three structural mainte-nance of chromosomes (SMC) complexes in eukaryotic cells

and contains Smc5, Smc6, and six non-SMC elements, Nse1–6(1–4). Similar to SMC proteins in the other two complexes,namely cohesin and condensin, Smc5 and Smc6 form the back-bone of the complex upon which Nse subunits assemble (3, 5, 6).However, the Smc5/6 complex contains a unique small ubiquitin-like modifier (SUMO) ligase subunit, Nse2/Mms21 (hereafterMms21), that facilitates the addition of SUMO to other proteins(1, 7, 8). The functions of the Smc5/6 complex are not as wellunderstood as those of cohesin and condensin. Nevertheless, theimportance of this complex is clear from the multiple defectsmanifested by its mutants, including sensitivity to DNA damage,telomere and rDNA defects, slow growth, and cell death (9, 10).Previous studies suggest that some of these defects indicate a rolefor the Smc5/6 complex in recombinational repair during im-paired replication. In particular, it was found that, in buddingyeast, when replicating in the presence of the DNA damagingagent methylmethane sulfonate (MMS), mutants of this complexaccumulate recombination intermediates, which are detected asX-shaped DNA molecules by 2D agarose gel electrophoresis (2Dgel) (11, 12). Because the absence of the DNA helicase Sgs1,which binds to Top3 and Rmi1 to dissolve double Hollidayjunctions (13), results in a similar defect, it was thought that theSmc5/6 complex might collaborate with Sgs1 to resolve suchrecombination intermediates (11, 14). However, a recent studysuggests that these two protein entities do not associate with eachother physically (12), raising the possibility that they performdifferent roles. Currently, the exact role of the Smc5/6 complexin recombinational repair is not known.

To understand the functions of the Smc5/6 complex, we soughtto identify proteins that interact physically with this complexusing yeast 2-hybrid (2H) screens coupled with co-immunopre-cipitation. From this study, we recovered the DNA helicaseMph1 as the only interactor that has known functions in recom-bination. Mph1 is not a core recombination protein; rather, itdefines a specialized recombination subpathway that operateswhen replication is impaired (15, 16). The nature of this sub-pathway remains to be elucidated but is expected to entail someof the biochemical activities exhibited by Mph1 and its orthologs,which include the Fanconi anemia M (FANCM) protein inhumans, the Fml proteins in fission yeast, and the archaeal Hefprotein (15, 17, 18). It is noteworthy that a FANCM mutation isimplicated in Fanconi anemia, a genetic disease that affectsdevelopment and causes cancer predisposition, suggesting theimportance of Mph1-like proteins in cell physiology (19, 20).

An activity shared by Mph1 and its orthologs is the dissocia-tion of DNA D-loop structures, a function important in limitingcrossovers during mitotic recombination (21–23). In addition,Fml1, FANCM, and Hef can catalyze the regression or unwind-ing of replication forks, which can potentially lead to recombi-nation (22–24). Although this process may subsequently restartthe damaged replication fork, it also increases the risk ofgenerating toxic recombination intermediates or other geneticalterations. Currently, it is unclear how this process is regulatedin vivo. In addition to recombination, Mph1 also has possibleroles in lagging-strand synthesis and in chromosomal rearrange-ment when overproduced (25, 26). Key residues for Mph1helicase activity are required only for recombinational repairand not for its other functions (15, 21, 25, 26).

We show here that binding to the Smc5/6 complex does notlead to Mph1 sumoylation. Rather, several major defects inmutants of the Smc5/6 complex are suppressed by mph1� andmph1 helicase mutations. These observed suppressions correlatewith a large reduction in the levels of recombination interme-diates seen in these mutants. In contrast, mph1� does notsuppress similar defects in sgs1� cells. Taken together, ourresults suggest that the Smc5/6 complex, but not Sgs1, modulatesthe Mph1-dependent recombination subpathway to prevent theaccumulation of recombination intermediates and the associateddeleterious effects.

ResultsMph1 Binds to the Smc5/6 Complex but Is Not Sumoylated. Tounderstand how the Smc5/6 complex is involved in recombina-

Author contributions: Y.-H.C., K.C., and X.Z. designed research; Y.-H.C., K.C., B.S., J.A., X.D.,and D.B. performed research; Y.-H.C. and K.C. contributed new reagents/analytic tools;Y.-H.C., K.C., B.S., X.D., H.Y., D.B., and X.Z. analyzed data; and X.Z. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

1Current address: University of Rochester Medical Center, Rochester, NY 14642.

2To whom correspondence should be addressed. E-mail: zhaox1@mskcc.org.

This article contains supporting information online at www.pnas.org/cgi/content/full/0908258106/DCSupplemental.

21252–21257 � PNAS � December 15, 2009 � vol. 106 � no. 50 www.pnas.org�cgi�doi�10.1073�pnas.0908258106

Dow

nloa

ded

by g

uest

on

Mar

ch 1

1, 2

021

tional repair, we examined whether it physically interacts withrecombination factors. Using each subunit of the Smc5/6 com-plex as bait, we performed 2H screens against an arrayed librarycontaining the majority of the yeast genes. The only interactingprotein with a known function in recombination recovered fromthis screen was the Mph1 helicase, with Smc5 as the bait. TheSmc5–Mph1 interaction was confirmed subsequently by pair-wise 2H tests (Fig. 1A).

We then constructed epitope-tagged Smc5 and Mph1 at theirchromosomal loci for co-immunoprecipitation experiments. Asshown in Fig. 1B, Mph1 pulled down Smc5, and vice versa (Fig.1B). In addition, Mph1 co-precipitated with Myc-tagged Smc6and Nse6 (Fig. 1 C and D). Furthermore, we found thatHis-tagged Smc5 protein expressed in Escherichia coli pulleddown recombinant Mph1 (Fig. 1E). These results demonstratethat Mph1 interacts with the Smc5/6 complex by directly bindingto Smc5.

Because the Mms21 subunit of the Smc5/6 complex is a SUMOE3 (1), the interaction of Mph1 with the Smc5/6 complex maysuggest that it is a SUMO target. However, we found that Mph1was not sumoylated during normal growth or after MMS treat-ment, when the sumoylation of several other recombinationproteins is induced (Fig. 1F). As a control, we can detect thesumoylation of Smc5 under both conditions (Fig. 1F). We thus

conclude that the Mph1–Smc5/6 complex interaction does notlead to Mph1 sumoylation.

MPH1 Deletion Suppresses Several Defects of Mutants of the Smc5/6Complex, Whereas Overexpression Exacerbates These Effects. Tounderstand the biological functions of the interaction betweenMph1 and the Smc5/6 complex, we examined the effects ofmph1� on three mutants of the Smc5/6 complex, includingsmc6–56 (27), smc6-P4 (or K239R), and mms21–11, which isdefective in sumoylation (1). All three alleles exhibit strongsensitivity to replication-blocking agents and slow growth, withsmc6–56 producing the most severe defects (Fig. 2 A and B). Wefound that mph1� suppressed the sensitivity to MMS, hydroxy-urea (HU), and UV of all three alleles (Fig. 2 A and B).Moreover, mph1� suppressed the slow growth of these mutants,an effect most evident in smc6–56 cells (Fig. 2 A and B). Inaddition to slow growth, smc6–56 cells also exhibit defectivecentromere separation (28). Strikingly, we found that mph1�restored centromere separation in smc6–56 cells from approx-imately 65% to above 90% (Fig. 2 C and D). We conclude thatthe deletion of MPH1 suppresses three major defects associatedwith mutants of the Smc5/6 complex, namely sensitivity to DNAdamage, slow growth, and defective centromere separation.

In a converse experiment, we found that MPH1 overexpres-sion severely inhibited the growth of smc6-P4 and mms21–11cells (Fig. 2E). To assess how MPH1 overexpression affectsMMS and HU sensitivities, we used endogenously YFP-taggedSmc6, which leads to normal growth but has mild defects in thecomplex’s functions as evidenced by its synthetic sick interactionwith mms21–11 (Fig. 2F). We found that MPH1 overexpressionrendered SMC6-YFP, but not WT cells, sensitive to MMS andHU (Fig. 2G). Taken together, the deletion and the overexpres-sion experiments show that Mph1 causes toxicity during normalgrowth and in the presence of genotoxins when the Smc5/6complex is defective.

These results prompted us to test whether Mph1 affects theessentiality of the Smc5/6 complex. Remarkably, we found thatdeletion of MPH1 suppressed the lethality of smc6� and mms21�spore clones (Fig. 2H). However, the double mutants mph1�smc6� and mph1� mms21� showed growth defects, indicatingthat the Smc5/6 complex possesses Mph1-independent functions.Nevertheless, the fact that cells can live without the Smc5/6complex if Mph1 is removed suggests that an important functionof the Smc5/6 complex pertains to Mph1 regulation duringnormal growth.

A Pro-Recombinogenic Function of Mph1 Is Toxic in Mutants of theSmc5/6 Complex. Mph1 has been shown to function in a subpath-way of recombinational repair to facilitate cell survival in MMS(15, 16). Consistent with previous reports, we found that mph1�exhibited moderate sensitivity to MMS, and very slight, if any,sensitivity to HU, UV, or X-rays and that mph1� did not increasethe DNA damage sensitivity of the recombination mutantsrad51� and rad52� (Fig. S1 A and B). In addition, we found thatMph1 formed nuclear foci during normal growth and that thepercentage of cells containing Mph1 foci increased moderatelyafter MMS treatment (Fig. 3A and Table S1). These Mph1 focifrequently co-localized with Rad52 and PCNA foci (Fig. 3A andTable S1), which are thought to represent recombination andreplication centers, respectively (29, 30). This cell biologicalresult is consistent with the genetic data and supports a role forMph1 in replication-associated recombinational repair.

Because Mph1 is also known to be involved in other processes,including lagging-strand synthesis and genomic rearrangements(25, 26), the question arises whether the toxicity of Mph1 inmutants of the Smc5/6 complex is related specifically to itsrecombination functions. We used two approaches to addressthis issue. First, we reasoned that if this hypothesis were true,

+

B

-Myc

-Flag

Mph1-FlagSmc5-Myc

++

+-

++ +

-IP: Myc IP: FlagInput Input

++

+-

++ +

-

Mph1

Smc5

C D

E

-Myc

-SUMO

Mph1Smc5

Smc5Mph1

SUMO-Smc5

Myc-tag:

FMph1-GST :Smc5-His6 :

++

+-

+-

++

-+

-

E W E W E W

stain 150

-GST

100

150 -

100 -

Mph1-GSTSmc5-His6

Mph1-GST

Mph1Smc6

A-T-L

-T-L-H-T-L-A

GAD-Mph1

GAD-vector

1 2 3 4 5 6 7GBD constructs

8

-T-L-T-L-H-T-L-A

Mph1-FlagNse6-Myc

IP: MycInput++

+-

++

+-

Mph1Nse6-Myc

-Flag-Myc-Flag

Mph1-FlagSmc6-Myc

IP: Myc++

+-

Input++

+-

Mph1Smc5-MMS +MMS

Fig. 1. Mph1 interacts with the Smc5/6 complex but is not sumoylated. (A)Mph1 interacts with Smc5 in 2H. The 2H strain pJ69–4a containing either aGAD-Mph1 or a GAD vector was mated to pJ69–4alpha cells containingGBD-fused Nse1–6 (lanes 1–6), Smc5 (lane 7), and Smc6 (lane 8). The resultingdiploids were selected on -TRP-LEU (-T-L) plates, and reporter activation wasscored by replica plating to -TRP-LEU-HIS (-T-L-H) and -TRP-LEU-ADE (-T-L-A)media. Note that GBD-Nse2/Mms21 self-activates the HIS3 and ADE2 report-ers. White boxes indicate the interactions between Mph1 and Smc5 and thecorresponding controls. (B–D) Mph1 co-immunoprecipitates with Smc5, Smc6,and Nse6. Lysates from cells containing the indicated tagged constructs wereimmunoprecipitated with the indicated antibody. Cell lysates (input) andimmunoprecipitated proteins (IP) were analyzed by protein blotting usinganti-Flag (Upper) and anti-Myc (Lower) antibodies. In (B), Smc5 was precipi-tated by anti-Flag antibody only in the presence of Mph1-Flag, and vice versa.In (C and D), Mph1 was precipitated by anti-Myc antibody only in the presenceof either Smc6-Myc (C) or Nse6-Myc (D). (E) Mph1 binds Smc5 in vitro. Recom-binant Mph1-GST and Smc5-His6 were expressed in E. coli, and Mph1-GST waspulled down by the Ni-NTA resins only in the presence of Smc5-His6. E: eluate,W: wash. (Top) Coomassie stained gel; (Bottom) Western blot using anti-GSTantibody. (F) Smc5, but not Mph1, is sumoylated. Cells containing Myc-taggedMph1 or Smc5 were grown in the absence (�MMS) or presence (�MMS) ofMMS. Mph1 and Smc5 were immunopurified by anti-Myc antibody and ex-amined by protein blotting using anti-Myc (Bottom) and anti-SUMO antibod-ies (Top).

Chen et al. PNAS � December 15, 2009 � vol. 106 � no. 50 � 21253

GEN

ETIC

S

Dow

nloa

ded

by g

uest

on

Mar

ch 1

1, 2

021

then eliminating recombination by means other than deletingMph1 should also suppress the defects in mutants of the Smc5/6complex. We found that the removal of the recombinase Rad51did indeed suppress the MMS and HU sensitivities of smc6–56and smc6-P4 cells to a degree similar to that observed for mph1�(Fig. 3B). However, rad51� did not significantly suppress theslow growth of smc6–56, suggesting that other recombinationalsubpathways may be required for the proper growth of thesecells.

In the second approach, we examined whether mutations ofkey residues of the Mph1 helicase domain can suppress defectsof smc6 and mms21 mutants, because Mph1 helicase activity isrequired for its roles in recombination but not for its otherfunctions (15, 21, 25, 26). To this end, we replaced WT MPH1with mph1-E210Q or mph1-Q603D, which mutated the con-served glutamic acid in the DEAH motif or the key residue ofthe helicase motif, respectively (15, 17). Consistent with previousreports, mph1-E210Q and -Q603D behaved like mph1� for DNAdamage sensitivity (Figs. 3C and S1A). We found that bothalleles suppressed the sensitivity to MMS, HU, and UV in smc6and mms21 mutant cells to the same degree as observed formph1� (Figs. 3D and S2 A and B). In addition, they suppressedthe slow growth of smc6–56 cells (Fig. 3D). Mph1-E210Q and

-Q603D proteins were expressed at levels similar to WT Mph1(Fig. 3E). Furthermore, Smc5 pulled down similar amounts ofMph1-Q603D and WT Mph1 (Fig. 3F), suggesting that Q603Ddoes not affect the Smc5–Mph1 interaction. Therefore weconclude that the suppression conferred by these mutations isnot caused by defects in protein levels or in the Mph1–Smc5interaction but rather by defects in the Mph1 helicase function.The results concerning rad51� and mph1 helicase mutationssuggest that a pro-recombinogenic role of Mph1 is toxic in cellswith defective Smc5/6 complexes.

Absence of Mph1 or Its Helicase Function Suppresses the Accumulationof Recombination Intermediates in smc6 and mms21 Mutants. Tounderstand the molecular basis of the suppression conferredby mph1� and mph1 helicase mutations, we performed 2D gelanalyses to examine the levels of recombination intermediatesproduced during impaired replication. Cells were synchro-nized in G2 phase and released into the cell cycle in thepresence of sublethal concentrations of MMS. DNA from WTcells and relevant mutants was extracted at different timepoints and examined in 2D gels using a probe for the earlyfiring replication origin ARS305 (Fig. 4A). It has been shownthat smc6–56 and mms21–11 cells, but not WT cells, accumu-

vec

B

mph1

smc6-P4

mph1 smc6-P4

WT

mms21-11

mph1 mms21-11

YPD MMS (0.001%) MMS (0.01%) HU (10 mM) UV (20 J/m2)A

mph1

smc6-56

mph1 smc6-56

WTYPD HU (10 mM) UV (40 J/m2)MMS (0.001%)

Glucose

mms21-11

WT

smc6-P4

vec

vec

Mph1

vec

Mph1

Mph1

GlucoseMMS HUNo drug

WT

SMC6-Y

E GGalactoseGalactose

smc6-56 smc6-56 mph10’30’60’90’120’

0 30 60 90 1200

20

40

60

80

100

% s

epar

atio

n

30 60 90 1200

20

40

60

80

100 mph1

WT

% s

epar

atio

n

WT mph10’30’60’90’120’

C D

min min

vec

Mph1

MMS HUNo drug

Mph1

smc6-56 mph1

smc6-56

0

F

mms21-

11mms2

1-11

SMC6-Y

SMC6-Y

WT

Hmms21 /+ mph1 /+ smc6 /+ mph1 /+

mms21mms21 mph1

smc6smc6 mph1

Fig. 2. mph1� rescues several defects of smc6 and mms21 mutants, whereas Mph1 overexpression confers opposite effects. (A and B) mph1� rescues the growthdefects and DNA damage sensitivity of smc6–56 (A), smc6-P4, and mms21–11 (B). WT and mutant cell cultures were diluted and spotted onto plates containingno drug or the indicated concentration of drugs or were treated with the indicated dose of UV light. (C and D) mph1� rescues centromere separation defectsin smc6–56 cells. Cells were arrested in G1 at 23 °C and were shifted to 37 °C for 1 h to inactivate the smc6–56 allele before being released from G1. Sampleswere collected every 15 min to examine cell-cycle progression by FACS and chromatid separation by the appearance of 2 dots of GFP-LacI bound to tandem LacOrepeats near the centromere on chromosome IV. To examine the events in the first cell cycle only, alpha-factor was added back to the cultures 45 min after release.(E and G) Mph1 overexpression sensitizes cells with defective Smc5/6 complexes. Cells containing smc6-P4 or mms21–11 (E) or SMC6-YFP (SMC6-Y, G) weretransformed with pGAL-Mph1 or the control vector; cell cultures were diluted and spotted on plates lacking uracil with glucose or galactose and with and withoutMMS or HU. (F) Smc6-YFP confers normal growth but is synthetic sick with mms21–11. A representative tetrad from diploid strains heterozygous for SMC6-YFPand mms21–11 is shown. The genotype for each spore clone is indicated. (H) mph1� suppresses the lethality of mms21� and smc6� cells. Representative tetradsfrom diploid strains with the indicated genotypes are shown. The spore clones containing mms21� mph1� (Left) and smc6� mph1� (Right) are underlined; thosecontaining mms21� (Left) and smc6� (Right) are marked with dotted lines, and their genotypes are deduced from sibling spore clones.

21254 � www.pnas.org�cgi�doi�10.1073�pnas.0908258106 Chen et al.

Dow

nloa

ded

by g

uest

on

Mar

ch 1

1, 2

021

late recombination intermediates at this DNA region and thatrad51� suppresses these defects (11, 12). We confirmed theseobservations and found that smc6-P4 cells also accumulatedX-shaped molecules for a prolonged period, whereas mph1�and mph1-Q603D behaved like WT cells (Fig. 4 A and B).Importantly, we found that both mph1� and mph1-Q603Dgreatly reduced the recombination intermediates detected in

smc6–56, smc6-P4, and mms21–11 cells (Figs. 4 A and B, andS2C). These results strongly suggest that the helicase activityof Mph1 is largely responsible for the accumulation of X-shaped recombination structures in smc6 and mms21 mutantcells during impaired replication. This observation provides alikely explanation for the rescue of smc6 and mms21 cells’sensitivity to MMS by mph1 mutations.

WT

DIC Mph1-YFP Rad52-RFP CFP-Pol30 MergeA

mph1rad51

WT

smc6-P4

YPD MMS (0.001%) HU (2 mM)(( ) (( )WT

smc6-56mph1 smc6-56

rad51

rad51 smc6-56

mph1

YPD MMS (0.001%)( HU (5 mM))))) ( ))B

mph1 smc6-P4rad51 smc6-P4

DC

E

WTsmc6-56

smc6-56 mph1-E210Qmph1-Q603D

smc6-56 mph1-Q603D

mph1-E210Q

YPD MMS (0.001%)(( )) HU (10 mM)(( )))))

F

WTYPD MMS (0.03%)

mph1mph1-E210Qmph1-Q603D

( ))

WT Q-D E-QMph1-YFP:

-YFP

stain

Mph1 Q-D

-YFP-Myc

Mph1-YFP:Smc5-Myc:

IP: MycQ-D

-WT+

WT-+

Smc5Mph1

InputQ-D

-WT+

WT-

Q-D+

Fig. 3. A pro-recombinogenic function of Mph1 is toxic in mutants of the Smc5/6 complex. (A) Mph1 foci co-localize with those of Rad52 and Pol30.Representative images of WT cells containing Mph1-YFP, Rad52-RFP, and CFP-Pol30 are shown. Foci formed by each protein are indicated by arrows. (B) rad51�,like mph1�, suppresses the sensitivity to MMS and HU of smc6-P4 cells (Left) and smc6–56 cells (Right). Experiments were performed as in Fig. 2A. (C and D)Helicase mutants mph1-E210Q and -Q603D exhibit sensitivity to MMS similar to that of mph1� (C) and suppress the sensitivity of smc6–56 cells to MMS and HU(D). Experiments were carried out as described in Fig. 2A. (E) Mph1-E210Q and -Q603D protein levels are similar to that of WT Mph1. WT and mutant Mph1proteins were tagged with YFP at their chromosomal loci; their protein levels were examined by protein blotting using anti-YFP antibody (Top). Similar loadingwas verified by amido-black staining (Bottom). (F) Similar amounts of WT Mph1 and Mph1-Q603D are immunoprecipitated with Smc5. Experiments were carriedout as in Fig. 1B.

A

smc6-P4

smc6-P4mph1

smc6-56

smc6-56mph1

logNoc60’120’180’

logNoc60’120’180’

logNoc60’120’180’

logNoc60’120’180’

WT

mms21-11

mph1

mms21-11mph1

60 120 180Min:

X moleculesLinear DNAY-arcs

Bubbles

ARS305HindIII

Chr III

EcoRV

Probe

B

WT

mph1-Q603D

smc6-56

smc6-56mph1-Q603D

60 120 180Min:

60 120 180Min:

logNoc60’120’180’

logNoc60’120’180’

logNoc60’120’180’

logNoc60’120’180’

logNoc60’120’180’

logNoc60’120’180’

logNoc60’120’180’

logNoc60’120’180’

Fig. 4. mph1� and mph1-Q603D decrease the levels of recombination intermediates in smc6 and mms21 mutants. (A and B) Cells were arrested at G2/M phasewith nocodazole at 25 °C and then were released into YPD medium with 0.033% MMS at 30 °C. The replication and recombination intermediates at the ARS305region 60, 120, and 180 min after release were analyzed by 2D gel electrophoresis followed by Southern blotting. Diagrams indicating the position of the probeand the replication structures are shown above the 2D gel images. The X-shaped DNA structures are indicated by arrowheads in smc6 and mms21 mutants. FACSanalyses are presented to the right of each gel image.

Chen et al. PNAS � December 15, 2009 � vol. 106 � no. 50 � 21255

GEN

ETIC

S

Dow

nloa

ded

by g

uest

on

Mar

ch 1

1, 2

021

The Genetic Interactions of Mph1 and the Smc5/6 Complex with theDNA Helicases Sgs1 and Srs2. Several defects of smc6 and mms21mutants are reminiscent of cells lacking the DNA helicase Sgs1.In particular, like smc6 and mms21 mutants, sgs1� cells accu-mulate recombination intermediates when cells replicate in thepresence of MMS, with rad51� suppressing this defect (14). Tounderstand whether mph1�, like rad51�, serves as a generalsuppressor for the accumulation of recombination structures, weexamined its effect in sgs1� cells. We found that under theexperimental conditions described above, the recombinationintermediates in sgs1� were largely unchanged for a prolongedperiod when MPH1 was deleted (Fig. 5A). In addition, mph1�did not suppress the sensitivity of sgs1� cells to MMS but insteadexacerbated it (Fig. 5B). Both observations are in contrast to thesuppression conferred by mph1� in smc6 and mms21 mutants,suggesting different roles of Sgs1 and the Smc5/6 complex inpreventing the accumulation of recombination intermediates.Consistent with this notion, sgs1� and mms21–11 are syntheticsick, and this defect is suppressed by the removal of Rad51(Fig. 5C).

Another DNA helicase that plays a role in the suppression ofrecombinational events is Srs2. A shared function of Srs2 andMph1 is preventing crossover events during mitotic recombina-tion (31–33). However, we found that, unlike mph1�, srs2�increased the sensitivity of mms21–11 cells to MMS (Fig. 5D). Inaddition, srs2� and mph1� were synergistic for sensitivity toMMS, suggesting that they perform different roles in cellularresistance to MMS (Fig. 5D). Taken together, these genetic dataindicate that the roles of the Smc5/6 complex in recombinationalrepair are at least partly different from those of Sgs1. Further-more, Mph1 has functions distinct from the other two helicases,Sgs1 and Srs2, with regard to cell survival in MMS.

DiscussionSeveral lines of genetic evidence have indicated the involvement ofthe Smc5/6 complex in recombinational repair, but the mechanismunderlying this function has been unclear. In this study, we haverevealed an interaction between the Smc5/6 complex and the DNAhelicase Mph1. We show that Mph1 is not sumoylated, although theSmc5/6 complex contains a SUMO ligase. Because mph1� allevi-ates several defects in smc6 and mms21 mutants, and MPH1

overexpression exacerbates some of these defects, Mph1 appears tobe toxic in these cells. This toxicity is caused by a pro-recombinogenic function of Mph1, because eliminating recombi-nation by rad51� or mutating key residues of Mph1 helicase motifs,required for recombination but not for other functions of Mph1,phenocopied mph1� in the suppression of smc6 and mms21 mu-tants. We further show that mph1� and its helicase mutations alsoreduced the levels of X-shaped recombination structures in smc6and mms21 mutant cells. Collectively, these results suggest that thehelicase activity of Mph1 leads to accumulation of recombinationintermediates that is detrimental in cells containing defectiveSmc5/6 complexes.

Our genetic and physical evidence, when combined with theknown biochemical activities of Mph1 and its orthologs, supportstwo models for the function of the Smc5/6 complex in recom-binational repair (Fig. 5E). The first model proposes that theSmc5/6 complex binds Mph1 and regulates its action as apro-recombinogenic factor. Mph1 orthologs can catalyze repli-cation fork regression/migration (22–24). It is thus possible thatthe Smc5/6 complex modulates Mph1 to prevent fork regressionor channels the regressed forks to direct restart instead ofrecombination. This action may prevent the formation of re-combination intermediates and permit the use of safer pathwaysfor rescuing damaged forks. The second model suggests that theSmc5/6 complex facilitates the resolution of the recombinationstructures generated by Mph1. Because Mph1 orthologs exhibitbranch migration activities, it is formally possible that Mph1 canpromote the formation of DNA joint molecules that need to beprocessed in an Smc5/6 complex-dependent manner. This rela-tionship would be somewhat similar to that of Sgs1 and Top3.However, because its subunits do not contain sequences char-acteristic of topoisomerases, helicases, or nucleases, the Smc5/6complex would play a structural role and/or recruit and assistadditional enzymes. The first model has more experimentalsupport at this stage. The Mph1 ortholog Fml1 has recently beenshown to promote recombination at stalled replication forks invivo, whereas it is unclear whether a potential role for Mph1 inbranch migration could lead to specific types of joint moleculesin cells. In fact, Mph1 was recently shown to dissociate D-loopstructures and to disfavor joint molecule formation in mitoticrecombination (21). Future work will be required to address the

mms21-11 rad51mms21-11 sgs1

A

mph1

sgs1

mph1sgs1

WTMMS (0.03%)YPDB

YPD MMS (0.01%)

mph1

WT

srs2

srs2 mph1

srs2 mms21-11

mms21-11

mms21-11

mms21-11/+ sgs1 /+ rad51 /+C

D

sgs1mph1

60 120 180Min: 240

logNoc60’120’180’240’

logNoc60’120’180’240’

Stalled RFs

RF regression

Model 1Stalled RFs & ssDNA gaps

Rad51-HR

Mph1Sgs1

Model 2

Smc5/6Linear DNA

Recombination intermediates

Top3Rad51-HR

Recombination intermediates

RF restart

E

sgs1

Mph1Smc5/6

( ))

mms21-11 rad51 sgs1

Fig. 5. Genetic interactions of mph1� and mms21–11 with sgs1� and srs2� and models for the functions of the Smc5/6 complex and Mph1 in recombinationalrepair. (A) The amount of X-shaped DNA in sgs1� cells is largely unaffected by mph1�. Experiments were carried out as in Fig. 4. (B and D) mph1� enhances theMMS sensitivity of sgs1� (B) and srs2� (D) cells, and srs2� exacerbates MMS sensitivity of mms21–11 cells (D). Experiments were carried out as in Fig. 2A, exceptthat plates were incubated for three days in B. (C) The synthetic sick interaction between sgs1� and mms21–11 is alleviated by rad51�. Shown are 3 tetrads fromindicated diploids; the spore clones of the relevant genotype are indicated. (E) Models for the functions of the Smc5/6 complex and Mph1 in recombinationalrepair. See text for details. RFs: replication forks, HR: homologous recombination.

21256 � www.pnas.org�cgi�doi�10.1073�pnas.0908258106 Chen et al.

Dow

nloa

ded

by g

uest

on

Mar

ch 1

1, 2

021

biochemical activities of Mph1 and the effect of the Smc5/6complex on such activities and on other aspects of Mph1functions.

In summary, the results presented here provide importantinsights into the role of the Smc5/6 complex in the Mph1-dependent recombination pathway. Because cells can live with-out the Smc5/6 complex when MPH1 is deleted, this role isimportant for normal growth. As defects in sgs1� cells are notsuppressed by mph1�, Sgs1 probably has limited roles in thispathway. rad51� suppression of the synthetic interaction of sgs1�and mms21–11 also supports the notion that Sgs1 and the Smc5/6complex have different roles in recombination. However, theSmc5/6 complex may work with proteins other than Sgs1 in theMph1 pathway, because mph1� was recently shown to suppressthe accumulation of X-shaped molecules in cells lacking Esc2, aprotein containing SUMO-like domains (34). Since the Smc5/6complex, Mph1, and Esc2 are conserved in humans, it will beinteresting to examine whether a similar recombination subpath-way exists in human cells. In this context, it is noteworthy that amutation in Mms21 leads to sensitivity to inter-strand cross-linking agents (35), a hallmark of Fanconi anemia cells. Becauseimproper recombination underlies the pathology of Fanconianemia and several other human diseases, an in-depth investi-gation of the mechanisms involved is likely to contribute to the

understanding and the development of potential treatments forthese conditions.

Materials and MethodsYeast Strains, Plasmids, Primers and Genetic Manipulations. Yeast strains andplasmids are listed in Table S2. Primer sequences are available upon request.We carried out 2H screens and pair-wise testing as described (6). Standardyeast protocols were used for strain construction, growth, and medium prep-aration. Spot assay plates were incubated at 30 °C and photographed aftertwo days, unless otherwise indicated.

Other Procedures. For live-cell imaging, cells were processed for microscopy asdescribed (29), except for the exposure times used for fusion proteins: CFP-Pol30, 0.3 s; Mph1-YFP, 3 s; and Rad52-RFP, 0.5 s. Spot assays for detectingsensitivity to damage and protein analyses were carried out as described (1).The primary antibodies used were anti-Myc (9E10), anti-Flag (Sigma), anti-GFP(Roche), and anti-SUMO (1). Chromatid separation assays (36), 2D gel analyses(11), and an in vitro protein pull-down assay (6) were performed as described.

ACKNOWLEDGMENTS. We thank S. Fields (University of Washington, Seattle,WA) for providing the 2H library; R. Rothstein (Columbia University, NewYork) and H. Klein (New York University, New York) for yeast strains; X. Liu forconstructing the smc6-P4 strain; Y. Yan for helping in sumoylation detection;and S. Keeney, M. Jasin, P. Thorpe, J. Petrini, and members of the Zhaolaboratory for comments on the manuscript. This work was supported by agrant from the Associazione Italiana per la Ricerca sul Cancro to D.B. and byNational Institute of Health Grants R01GM079516 to H.Y. and R01GM080670to X.Z.

1. Zhao X, Blobel G (2005) A SUMO ligase is part of a nuclear multiprotein complex thataffects DNA repair and chromosomal organization. Proc Natl Acad Sci USA 102:4777–4782.

2. Sergeant J, et al. (2005) Composition and architecture of the Schizosaccharomycespombe Rad18 (Smc5–6) complex. Mol Cell Biol 25:172–184.

3. Pebernard S, Wohlschlegel J, McDonald WH, Yates JR, III, Boddy MN (2006) TheNse5-Nse6 dimer mediates DNA repair roles of the Smc5-Smc6 complex. Mol Cell Biol26:1617–1630.

4. Taylor EM, Copsey AC, Hudson JJ, Vidot S, Lehmann AR (2008) Identification of theproteins, including MAGEG1, that make up the human SMC5–6 protein complex. MolCell Biol 28:1197–1206.

5. Palecek J, Vidot S, Feng M, Doherty AJ, Lehmann AR (2006) The Smc5-Smc6 DNA repaircomplex. Bridging of the Smc5-Smc6 heads by the KLEISIN, Nse4, and non-Kleisinsubunits. J Biol Chem 281:36952–36959.

6. Duan X, et al. (2009) The architecture of the Smc5/6 complex of S. cerevisiae reveals aunique interaction between the Nse5–6 subcomplex and the hinge regions of Smc5and Smc6. J Biol Chem 284:8507–8515.

7. Andrews EA, et al. (2005) Nse2, a component of the Smc5–6 complex, is a SUMO ligaserequired for the response to DNA damage. Mol Cell Biol 25:185–196.

8. Potts PR, Yu H (2005) Human MMS21/NSE2 is a SUMO ligase required for DNA repair.Mol Cell Biol 25:7021–7032.

9. Murray JM, Carr AM (2008) Smc5/6: A link between DNA repair and unidirectionalreplication? Nat Rev Mol Cell Biol 9:177–182.

10. De Piccoli G, Torres-Rosell J, Aragon L (2009) The unnamed complex: What do we knowabout Smc5-Smc6? Chromosome Res 17:251–263.

11. Branzei D, et al. (2006) Ubc9- and Mms21-mediated sumoylation counteracts recom-binogenic events at damaged replication forks. Cell 127:509–522.

12. Sollier J, et al. (2009) The S. cerevisiae Esc2 and Smc5–6 proteins promote sisterchromatid junction mediated intra-S repair. Mol Biol Cell 20:1671–1682.

13. Mankouri HW, Hickson ID (2007) The RecQ helicase-topoisomerase III-Rmi1 complex:ADNA structure-specific ‘dissolvasome’? Trends Biochem Sci 32:538–546.

14. Liberi G, et al. (2005) Rad51-dependent DNA structures accumulate at damagedreplication forks in sgs1 mutants defective in the yeast ortholog of BLM RecQ helicase.Genes Dev 19:339–350.

15. Schurer KA, Rudolph C, Ulrich HD, Kramer W (2004) Yeast MPH1 gene functions in anerror-free DNA damage bypass pathway that requires genes from homologous recom-bination, but not from postreplicative repair. Genetics 166:1673–1686.

16. St Onge RP, et al. (2007) Systematic pathway analysis using high-resolution fitnessprofiling of combinatorial gene deletions. Nat Genet 39:199–206.

17. Prakash R, et al. (2005) S. cerevisiae MPH1 gene, required for homologous recombi-nation-mediated mutation avoidance, encodes a 3� to 5� DNA helicase. J Biol Chem280:7854–7860.

18. Nishino T, Komori K, Tsuchiya D, Ishino Y, Morikawa K (2005) Crystal structure andfunctional implications of Pyrococcus furiosus hef helicase domain involved inbranched DNA processing. Structure (London) 13:143–153.

19. Meetei AR, et al. (2005) A human ortholog of archaeal DNA repair protein Hef isdefective in Fanconi anemia complementation group M. Nat Genet 37:958–963.

20. Mosedale G, et al. (2005) The vertebrate Hef ortholog is a component of the Fanconianemia tumor-suppressor pathway. Nature Structural and Molecular Biology 12:763–771.

21. Prakash R, et al. (2009) Yeast Mph1 helicase dissociates Rad51-made D-loops: Implica-tions for crossover control in mitotic recombination. Genes Dev 23:67–79.

22. Sun W, et al. (2008) The FANCM ortholog Fml1 promotes recombination at stalledreplication forks and limits crossing over during DNA double-strand break repair. MolCell 32:118–128.

23. Gari K, Decaillet C, Stasiak AZ, Stasiak A, Constantinou A (2008) The Fanconi anemiaprotein FANCM can promote branch migration of Holliday junctions and replicationforks. Mol Cell 29:141–148.

24. Komori K, et al. (2004) Cooperation of the N-terminal Helicase and C-terminal endo-nuclease activities of archaeal Hef protein in processing stalled replication forks. J BiolChem 279:53175–53185.

25. Banerjee S, et al. (2008) Mph1p promotes gross chromosomal rearrangement throughpartial inhibition of homologous recombination. J Cell Biol 181:1083–1093.

26. Kang YH, et al. (2009) The MPH1 gene of S. cerevisiae functions in Okazaki fragmentsprocessing. J Biol Chem 284:10376–10386.

27. Onoda F, et al. (2004) SMC6 is required for MMS-induced interchromosomal and sisterchromatid recombinations in S. cerevisiae. DNA Repair 3:429–439.

28. Lindroos HB, et al. (2006) Chromosomal association of the Smc5/6 complex reveals thatit functions in differently regulated pathways. Mol Cell 22:755–767.

29. Lisby M, Barlow JH, Burgess RC, Rothstein R (2004) Choreography of the DNA damageresponse: Spatiotemporal relationships among checkpoint and repair proteins. Cell118:699–713.

30. Kitamura E, Blow JJ, Tanaka TU (2006) Live-cell imaging reveals replication of individ-ual replicons in eukaryotic replication factories. Cell 125:1297–1308.

31. Ira G, Malkova A, Liberi G, Foiani M, Haber JE (2003) Srs2 and Sgs1-Top3 suppresscrossovers during double-strand break repair in yeast. Cell 115:401–411.

32. Lo YC, et al. (2006) Sgs1 regulates gene conversion tract lengths and crossoversindependently of its helicase activity. Mol Cell Biol 26:4086–4094.

33. Robert T, Dervins D, Fabre F, Gangloff S (2006) Mrc1 and Srs2 are major actors in theregulation of spontaneous crossover. EMBO J 25:2837–2846.

34. Mankouri HW, NgoHP, Hickson ID (2009) Esc2 and Sgs1 act in functionally distinctbranches of the homologous recombination repair pathway in S. cerevisiae. Mol BiolCell 20:1683–1694.

35. Hoch NC, et al. (2008) Allelism of Saccharomyces cerevisiae gene PSO10, involved inerror-prone repair of psoralen-induced DNA damage, with SUMO ligase-encodingMMS21. Curr Genet 53:361–371.

36. Straight AF, Belmont AS, Robinett CC, Murray AW (1996) GFP tagging of budding yeastchromosomes reveals that protein–protein interactions can mediate sister chromatidcohesion. Curr Biol 6:1599–1608.

Chen et al. PNAS � December 15, 2009 � vol. 106 � no. 50 � 21257

GEN

ETIC

S

Dow

nloa

ded

by g

uest

on

Mar

ch 1

1, 2

021